The long term goal of this project is to develop methods that will allow researchers to gain insight into the metabolic networks of organisms for which we have little or no high-throughput data. Such metabolic networks can reveal aspects of the organism's metabolism that might make it vulnerable to new or existing therapies. A core data set using genomic and other omic data from data-rich bacteria that are related to the organisms of interest will be assembled. The statistical tools needed to integrate these data and to infer metabolic networks using these core data plus characterization (phenotypic) data will then be built. Using the statistical inference algorithms, the characterization data can be leveraged to reveal the metabolic networks of data-poor bacteria for which we have only characterization data. This approach can eliminate the need for genome sequencing, gene expression experiments and the like for thousands of Gram-negative facultative rod bacteria (GNF). There are five tasks in the project: (1) assemble the data sets from data-rich organisms that will be used to inform the inference algorithm. These data include (a) the genomic sequences and annotation information, (b) extant pathway data and (c) gene expression data. All these data contain some level of information about the connectivity within the metabolic network; (2) process the genomic data to enhance its predictive value; (3) develop a data integration algorithm; (4) investigate modeling frameworks to be used for Bayesian data fusion and network inference; (5) validate the metabolic networks. Deliverables from this project should include: (1) a set of pathway genome databases for 35 GNF, This group includes 20 strains classified as category A or B biothreat agents, (2) a core dataset that integrates all the information we have relevant to the metabolic pathways in the 35 sequenced GNF, (3) a probabilistic graphical modeling framework capable of integrating disparate types of data and inferring networks from the integrated data, (4) a method for using characterization data, along with deliverables 2 and 3, to infer metabolic networks for bacterial strains for which we have only characterization data. The ability to rapidly construct models of metabolic networks means researchers will be able to respond to emerging infectious agents or biothreats more quickly. Relevance The methods developed as part of this proposal will allow us to quickly make metabolic maps for thousands of bacteria. Such maps can guide researchers to promising new targets for therapeutic or preventative measures against pathogenic bacteria. The fight against well-known pathogens and biothreat agents, as well as against new, emerging pathogens will be greatly aided by these tools. [unreadable] [unreadable] [unreadable] [unreadable]